lorinachey/rad-rl-checkpoints

Boston Dynamics Spot Flat Terrain Locomotion Policy

Model Description

A low-level locomotion policy for Boston Dynamics Spot trained via Proximal Policy Optimization (PPO) in Isaac Lab. The policy takes proprioceptive state observations and outputs joint-level actions to achieve stable locomotion on flat terrain.

Training Details

Parameter	Value
Framework	Isaac Lab
Training Library	RSL-RL
Algorithm	PPO
Environment	Spot flat terrain
Checkpoint	model_19999.pt
Training Iterations	20,000

Observations and Actions

• Observations: Proprioceptive state (joint positions, joint velocities, base linear velocity, base angular velocity, projected gravity vector, velocity commands)
• Actions: Target joint positions for all 12 Spot joints (3 per leg × 4 legs)

Robot

• Platform: Boston Dynamics Spot (quadruped, 12 DoF)

• Simulation: Isaac Lab (Isaac Sim)

• Developed by: [More Information Needed]

• Funded by [optional]: [More Information Needed]

• Shared by [optional]: [More Information Needed]

• Model type: [More Information Needed]

• Language(s) (NLP): [More Information Needed]

• License: mit

• Finetuned from model [optional]: [More Information Needed]

Model Sources [optional]

• Repository: https://github.com/leggedrobotics/rsl_rl
• Paper: https://arxiv.org/abs/2509.10771

Citation

If you use this model, please cite the RSL-RL library used for training:

@article{schwarke2025rslrl,
  title={RSL-RL: A Learning Library for Robotics Research},
  author={Schwarke, Clemens and Mittal, Mayank and Rudin, Nikita and Hoeller, David and Hutter, Marco},
  journal={arXiv preprint arXiv:2509.10771},
  year={2025}
}

Note the paper is a 2025 arXiv preprint — it's available at arXiv:2509.10771.

Model Card Authors [optional]

• Lorin Achey

Model Card Contact

• loac4399@colorado.edu